Dear Mohamed,
what you influence with WFAC1 is (in XDS terminology) the number of rejected "misfits", _not_ "aliens".
Remember that rejection of outliers is meant to improve the accuracy, not just the precision (which is what R values, I/sigma and cc1/2 measure). Precision always increases when rejecting more outliers, but accuracy may degrade! A good rule is that there should be a verified reason if more than the usual fraction of up to 1% should be rejected: ice rings, shadow on detector, ... You should run XDSSTAT and inspect misfits.pck, rf.pck in addition to FRAME. cbf.
It is ok to increase WFAC1 but I have never seen a reason to decrease below 1.
HTH,
Kay
On Tue, 18 Aug 2015 23:09:56 +0100, Mohamed Noor <[log in to unmask]> wrote:
>Dear all
>
>I note from the XDS manual and wiki that the WFAC1 keyword, IIUC, can be used to control the number of 'alien' reflections. How does one differentiate between a true outlier (ice?) and those that should be included in the final dataset instead of being removed?
>
>The datasets in question were collected at the peak wavelength of anomalous scatterer and have about 1000-2000 aliens with WFAC1=1 and reduced to about 200 with WFAC1=1.5.
>
>What is the downside of increasing the WFAC1? Can it be increased to say, 2?
>
>Thanks.
|